PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
March 25, 2026The Journal of SupercomputingOpen Access

Accelerated deep learning denoising for edge AI in single-pixel imaging

View Full Paper
Ask AI
Bookmark
Share

Authors

CCCarlos Chabert-UllHTHeberley Tobón-MayaSZSamuel I. Zapata-Valencia

Discussion

Loading...

Member takes

Overview

Demonstrates efficient image reconstruction in single-pixel imaging using a deep learning approach on edge AI hardware, suggesting improved real-time applications.

Key Points

  • The aim is to develop a real-time single-pixel imaging system using deep learning denoising on embedded GPU hardware.
  • Utilized compressed sensing with Hadamard patterns for image reconstruction.
  • Trained a compact U-Net model using mean squared error on simulated grayscale faces.
  • Deployed the model on a Jetson Orin NX 16 GB for efficient processing.
  • Evaluated performance based on acquisition, reconstruction, and inference times.
  • Assessed quality using PSNR and SSIM metrics.
  • GPU inference significantly reduces denoising time.
  • The bottleneck shifts towards optical acquisition as sampling increases.
  • The proposed methodology is reproducible for future SPI denoising assessments.
  • Identified strategies to enhance throughput such as higher-rate pattern projection.

Cite This Study

Chabert-Ull et al. (2026) studied this question.

synapsesocial.com/papers/69c37bf3b34aaaeb1a67eca1https://doi.org/10.1007/s11227-026-08391-y
View Full Paper
Ask AI
Bookmark
Share